1 |
root |
1.1 |
=head1 NAME |
2 |
|
|
|
3 |
root |
1.89 |
AnyEvent::MP - erlang-style multi-processing/message-passing framework |
4 |
root |
1.1 |
|
5 |
|
|
=head1 SYNOPSIS |
6 |
|
|
|
7 |
|
|
use AnyEvent::MP; |
8 |
|
|
|
9 |
root |
1.75 |
$NODE # contains this node's node ID |
10 |
|
|
NODE # returns this node's node ID |
11 |
root |
1.2 |
|
12 |
root |
1.38 |
$SELF # receiving/own port id in rcv callbacks |
13 |
|
|
|
14 |
root |
1.48 |
# initialise the node so it can send/receive messages |
15 |
root |
1.72 |
configure; |
16 |
root |
1.48 |
|
17 |
root |
1.75 |
# ports are message destinations |
18 |
root |
1.38 |
|
19 |
|
|
# sending messages |
20 |
root |
1.2 |
snd $port, type => data...; |
21 |
root |
1.38 |
snd $port, @msg; |
22 |
|
|
snd @msg_with_first_element_being_a_port; |
23 |
root |
1.2 |
|
24 |
root |
1.50 |
# creating/using ports, the simple way |
25 |
root |
1.73 |
my $simple_port = port { my @msg = @_ }; |
26 |
root |
1.22 |
|
27 |
root |
1.52 |
# creating/using ports, tagged message matching |
28 |
root |
1.38 |
my $port = port; |
29 |
root |
1.73 |
rcv $port, ping => sub { snd $_[0], "pong" }; |
30 |
|
|
rcv $port, pong => sub { warn "pong received\n" }; |
31 |
root |
1.2 |
|
32 |
root |
1.48 |
# create a port on another node |
33 |
|
|
my $port = spawn $node, $initfunc, @initdata; |
34 |
|
|
|
35 |
root |
1.116 |
# destroy a port again |
36 |
root |
1.101 |
kil $port; # "normal" kill |
37 |
|
|
kil $port, my_error => "everything is broken"; # error kill |
38 |
|
|
|
39 |
root |
1.35 |
# monitoring |
40 |
root |
1.90 |
mon $localport, $cb->(@msg) # callback is invoked on death |
41 |
|
|
mon $localport, $otherport # kill otherport on abnormal death |
42 |
|
|
mon $localport, $otherport, @msg # send message on death |
43 |
root |
1.35 |
|
44 |
root |
1.101 |
# temporarily execute code in port context |
45 |
|
|
peval $port, sub { die "kill the port!" }; |
46 |
|
|
|
47 |
|
|
# execute callbacks in $SELF port context |
48 |
|
|
my $timer = AE::timer 1, 0, psub { |
49 |
|
|
die "kill the port, delayed"; |
50 |
|
|
}; |
51 |
|
|
|
52 |
root |
1.45 |
=head1 CURRENT STATUS |
53 |
|
|
|
54 |
root |
1.71 |
bin/aemp - stable. |
55 |
|
|
AnyEvent::MP - stable API, should work. |
56 |
elmex |
1.77 |
AnyEvent::MP::Intro - explains most concepts. |
57 |
root |
1.88 |
AnyEvent::MP::Kernel - mostly stable API. |
58 |
|
|
AnyEvent::MP::Global - stable API. |
59 |
root |
1.45 |
|
60 |
root |
1.1 |
=head1 DESCRIPTION |
61 |
|
|
|
62 |
root |
1.2 |
This module (-family) implements a simple message passing framework. |
63 |
|
|
|
64 |
|
|
Despite its simplicity, you can securely message other processes running |
65 |
root |
1.67 |
on the same or other hosts, and you can supervise entities remotely. |
66 |
root |
1.2 |
|
67 |
root |
1.23 |
For an introduction to this module family, see the L<AnyEvent::MP::Intro> |
68 |
root |
1.67 |
manual page and the examples under F<eg/>. |
69 |
root |
1.23 |
|
70 |
root |
1.2 |
=head1 CONCEPTS |
71 |
|
|
|
72 |
|
|
=over 4 |
73 |
|
|
|
74 |
|
|
=item port |
75 |
|
|
|
76 |
root |
1.79 |
Not to be confused with a TCP port, a "port" is something you can send |
77 |
|
|
messages to (with the C<snd> function). |
78 |
root |
1.29 |
|
79 |
root |
1.53 |
Ports allow you to register C<rcv> handlers that can match all or just |
80 |
root |
1.64 |
some messages. Messages send to ports will not be queued, regardless of |
81 |
|
|
anything was listening for them or not. |
82 |
root |
1.2 |
|
83 |
root |
1.119 |
Ports are represented by (printable) strings called "port IDs". |
84 |
|
|
|
85 |
root |
1.67 |
=item port ID - C<nodeid#portname> |
86 |
root |
1.2 |
|
87 |
root |
1.123 |
A port ID is the concatenation of a node ID, a hash-mark (C<#>) |
88 |
|
|
as separator, and a port name (a printable string of unspecified |
89 |
|
|
format created by AnyEvent::MP). |
90 |
root |
1.2 |
|
91 |
|
|
=item node |
92 |
|
|
|
93 |
root |
1.53 |
A node is a single process containing at least one port - the node port, |
94 |
root |
1.67 |
which enables nodes to manage each other remotely, and to create new |
95 |
root |
1.53 |
ports. |
96 |
root |
1.2 |
|
97 |
root |
1.67 |
Nodes are either public (have one or more listening ports) or private |
98 |
|
|
(no listening ports). Private nodes cannot talk to other private nodes |
99 |
root |
1.119 |
currently, but all nodes can talk to public nodes. |
100 |
|
|
|
101 |
|
|
Nodes is represented by (printable) strings called "node IDs". |
102 |
root |
1.2 |
|
103 |
root |
1.117 |
=item node ID - C<[A-Za-z0-9_\-.:]*> |
104 |
root |
1.2 |
|
105 |
root |
1.64 |
A node ID is a string that uniquely identifies the node within a |
106 |
|
|
network. Depending on the configuration used, node IDs can look like a |
107 |
|
|
hostname, a hostname and a port, or a random string. AnyEvent::MP itself |
108 |
root |
1.119 |
doesn't interpret node IDs in any way except to uniquely identify a node. |
109 |
root |
1.64 |
|
110 |
|
|
=item binds - C<ip:port> |
111 |
|
|
|
112 |
|
|
Nodes can only talk to each other by creating some kind of connection to |
113 |
|
|
each other. To do this, nodes should listen on one or more local transport |
114 |
root |
1.119 |
endpoints - binds. |
115 |
|
|
|
116 |
|
|
Currently, only standard C<ip:port> specifications can be used, which |
117 |
|
|
specify TCP ports to listen on. So a bind is basically just a tcp socket |
118 |
|
|
in listening mode thta accepts conenctions form other nodes. |
119 |
root |
1.64 |
|
120 |
root |
1.83 |
=item seed nodes |
121 |
root |
1.64 |
|
122 |
root |
1.119 |
When a node starts, it knows nothing about the network it is in - it |
123 |
|
|
needs to connect to at least one other node that is already in the |
124 |
|
|
network. These other nodes are called "seed nodes". |
125 |
|
|
|
126 |
|
|
Seed nodes themselves are not special - they are seed nodes only because |
127 |
|
|
some other node I<uses> them as such, but any node can be used as seed |
128 |
|
|
node for other nodes, and eahc node cna use a different set of seed nodes. |
129 |
root |
1.83 |
|
130 |
|
|
In addition to discovering the network, seed nodes are also used to |
131 |
root |
1.119 |
maintain the network - all nodes using the same seed node form are part of |
132 |
|
|
the same network. If a network is split into multiple subnets because e.g. |
133 |
|
|
the network link between the parts goes down, then using the same seed |
134 |
|
|
nodes for all nodes ensures that eventually the subnets get merged again. |
135 |
root |
1.83 |
|
136 |
|
|
Seed nodes are expected to be long-running, and at least one seed node |
137 |
root |
1.85 |
should always be available. They should also be relatively responsive - a |
138 |
|
|
seed node that blocks for long periods will slow down everybody else. |
139 |
root |
1.83 |
|
140 |
root |
1.119 |
For small networks, it's best if every node uses the same set of seed |
141 |
|
|
nodes. For large networks, it can be useful to specify "regional" seed |
142 |
|
|
nodes for most nodes in an area, and use all seed nodes as seed nodes for |
143 |
|
|
each other. What's important is that all seed nodes connections form a |
144 |
|
|
complete graph, so that the network cannot split into separate subnets |
145 |
|
|
forever. |
146 |
|
|
|
147 |
|
|
Seed nodes are represented by seed IDs. |
148 |
|
|
|
149 |
|
|
=item seed IDs - C<host:port> |
150 |
root |
1.83 |
|
151 |
root |
1.119 |
Seed IDs are transport endpoint(s) (usually a hostname/IP address and a |
152 |
elmex |
1.96 |
TCP port) of nodes that should be used as seed nodes. |
153 |
root |
1.29 |
|
154 |
root |
1.119 |
=item global nodes |
155 |
|
|
|
156 |
|
|
An AEMP network needs a discovery service - nodes need to know how to |
157 |
|
|
connect to other nodes they only know by name. In addition, AEMP offers a |
158 |
|
|
distributed "group database", which maps group names to a list of strings |
159 |
|
|
- for example, to register worker ports. |
160 |
|
|
|
161 |
|
|
A network needs at least one global node to work, and allows every node to |
162 |
|
|
be a global node. |
163 |
|
|
|
164 |
|
|
Any node that loads the L<AnyEvent::MP::Global> module becomes a global |
165 |
|
|
node and tries to keep connections to all other nodes. So while it can |
166 |
|
|
make sense to make every node "global" in small networks, it usually makes |
167 |
|
|
sense to only make seed nodes into global nodes in large networks (nodes |
168 |
|
|
keep connections to seed nodes and global nodes, so makign them the same |
169 |
|
|
reduces overhead). |
170 |
root |
1.67 |
|
171 |
root |
1.2 |
=back |
172 |
|
|
|
173 |
root |
1.3 |
=head1 VARIABLES/FUNCTIONS |
174 |
root |
1.2 |
|
175 |
|
|
=over 4 |
176 |
|
|
|
177 |
root |
1.1 |
=cut |
178 |
|
|
|
179 |
|
|
package AnyEvent::MP; |
180 |
|
|
|
181 |
root |
1.121 |
use AnyEvent::MP::Config (); |
182 |
root |
1.44 |
use AnyEvent::MP::Kernel; |
183 |
root |
1.121 |
use AnyEvent::MP::Kernel qw(%NODE %PORT %PORT_DATA $UNIQ $RUNIQ $ID); |
184 |
root |
1.2 |
|
185 |
root |
1.1 |
use common::sense; |
186 |
|
|
|
187 |
root |
1.2 |
use Carp (); |
188 |
|
|
|
189 |
root |
1.1 |
use AE (); |
190 |
root |
1.124 |
use Guard (); |
191 |
root |
1.1 |
|
192 |
root |
1.2 |
use base "Exporter"; |
193 |
|
|
|
194 |
root |
1.121 |
our $VERSION = $AnyEvent::MP::Config::VERSION; |
195 |
root |
1.43 |
|
196 |
root |
1.8 |
our @EXPORT = qw( |
197 |
root |
1.59 |
NODE $NODE *SELF node_of after |
198 |
root |
1.72 |
configure |
199 |
root |
1.101 |
snd rcv mon mon_guard kil psub peval spawn cal |
200 |
root |
1.22 |
port |
201 |
root |
1.124 |
db_set db_del db_reg |
202 |
root |
1.8 |
); |
203 |
root |
1.2 |
|
204 |
root |
1.22 |
our $SELF; |
205 |
|
|
|
206 |
|
|
sub _self_die() { |
207 |
|
|
my $msg = $@; |
208 |
|
|
$msg =~ s/\n+$// unless ref $msg; |
209 |
|
|
kil $SELF, die => $msg; |
210 |
|
|
} |
211 |
|
|
|
212 |
|
|
=item $thisnode = NODE / $NODE |
213 |
|
|
|
214 |
root |
1.67 |
The C<NODE> function returns, and the C<$NODE> variable contains, the node |
215 |
root |
1.64 |
ID of the node running in the current process. This value is initialised by |
216 |
root |
1.72 |
a call to C<configure>. |
217 |
root |
1.22 |
|
218 |
root |
1.63 |
=item $nodeid = node_of $port |
219 |
root |
1.22 |
|
220 |
root |
1.67 |
Extracts and returns the node ID from a port ID or a node ID. |
221 |
root |
1.34 |
|
222 |
root |
1.78 |
=item configure $profile, key => value... |
223 |
|
|
|
224 |
root |
1.72 |
=item configure key => value... |
225 |
root |
1.34 |
|
226 |
root |
1.64 |
Before a node can talk to other nodes on the network (i.e. enter |
227 |
root |
1.72 |
"distributed mode") it has to configure itself - the minimum a node needs |
228 |
root |
1.64 |
to know is its own name, and optionally it should know the addresses of |
229 |
|
|
some other nodes in the network to discover other nodes. |
230 |
root |
1.34 |
|
231 |
root |
1.121 |
This function configures a node - it must be called exactly once (or |
232 |
|
|
never) before calling other AnyEvent::MP functions. |
233 |
|
|
|
234 |
root |
1.108 |
The key/value pairs are basically the same ones as documented for the |
235 |
root |
1.121 |
F<aemp> command line utility (sans the set/del prefix), with two additions: |
236 |
|
|
|
237 |
|
|
=over 4 |
238 |
|
|
|
239 |
|
|
=item norc => $boolean (default false) |
240 |
|
|
|
241 |
|
|
If true, then the rc file (e.g. F<~/.perl-anyevent-mp>) will I<not> |
242 |
|
|
be consulted - all configuraiton options must be specified in the |
243 |
|
|
C<configure> call. |
244 |
root |
1.108 |
|
245 |
root |
1.121 |
=item force => $boolean (default false) |
246 |
|
|
|
247 |
|
|
IF true, then the values specified in the C<configure> will take |
248 |
|
|
precedence over any values configured via the rc file. The default is for |
249 |
|
|
the rc file to override any options specified in the program. |
250 |
|
|
|
251 |
|
|
=back |
252 |
root |
1.34 |
|
253 |
root |
1.72 |
=over 4 |
254 |
|
|
|
255 |
|
|
=item step 1, gathering configuration from profiles |
256 |
|
|
|
257 |
|
|
The function first looks up a profile in the aemp configuration (see the |
258 |
|
|
L<aemp> commandline utility). The profile name can be specified via the |
259 |
root |
1.78 |
named C<profile> parameter or can simply be the first parameter). If it is |
260 |
|
|
missing, then the nodename (F<uname -n>) will be used as profile name. |
261 |
root |
1.34 |
|
262 |
root |
1.72 |
The profile data is then gathered as follows: |
263 |
root |
1.69 |
|
264 |
elmex |
1.77 |
First, all remaining key => value pairs (all of which are conveniently |
265 |
root |
1.72 |
undocumented at the moment) will be interpreted as configuration |
266 |
|
|
data. Then they will be overwritten by any values specified in the global |
267 |
|
|
default configuration (see the F<aemp> utility), then the chain of |
268 |
|
|
profiles chosen by the profile name (and any C<parent> attributes). |
269 |
|
|
|
270 |
|
|
That means that the values specified in the profile have highest priority |
271 |
|
|
and the values specified directly via C<configure> have lowest priority, |
272 |
|
|
and can only be used to specify defaults. |
273 |
root |
1.49 |
|
274 |
root |
1.64 |
If the profile specifies a node ID, then this will become the node ID of |
275 |
root |
1.122 |
this process. If not, then the profile name will be used as node ID, with |
276 |
|
|
a slash (C</>) attached. |
277 |
|
|
|
278 |
|
|
If the node ID (or profile name) ends with a slash (C</>), then a random |
279 |
|
|
string is appended to make it unique. |
280 |
root |
1.64 |
|
281 |
root |
1.72 |
=item step 2, bind listener sockets |
282 |
|
|
|
283 |
root |
1.64 |
The next step is to look up the binds in the profile, followed by binding |
284 |
|
|
aemp protocol listeners on all binds specified (it is possible and valid |
285 |
|
|
to have no binds, meaning that the node cannot be contacted form the |
286 |
|
|
outside. This means the node cannot talk to other nodes that also have no |
287 |
|
|
binds, but it can still talk to all "normal" nodes). |
288 |
|
|
|
289 |
root |
1.70 |
If the profile does not specify a binds list, then a default of C<*> is |
290 |
root |
1.72 |
used, meaning the node will bind on a dynamically-assigned port on every |
291 |
|
|
local IP address it finds. |
292 |
|
|
|
293 |
|
|
=item step 3, connect to seed nodes |
294 |
root |
1.64 |
|
295 |
root |
1.119 |
As the last step, the seed ID list from the profile is passed to the |
296 |
root |
1.64 |
L<AnyEvent::MP::Global> module, which will then use it to keep |
297 |
root |
1.72 |
connectivity with at least one node at any point in time. |
298 |
root |
1.64 |
|
299 |
root |
1.72 |
=back |
300 |
|
|
|
301 |
root |
1.87 |
Example: become a distributed node using the local node name as profile. |
302 |
root |
1.72 |
This should be the most common form of invocation for "daemon"-type nodes. |
303 |
root |
1.34 |
|
304 |
root |
1.72 |
configure |
305 |
root |
1.34 |
|
306 |
root |
1.64 |
Example: become an anonymous node. This form is often used for commandline |
307 |
|
|
clients. |
308 |
root |
1.34 |
|
309 |
root |
1.72 |
configure nodeid => "anon/"; |
310 |
|
|
|
311 |
root |
1.120 |
Example: configure a node using a profile called seed, which is suitable |
312 |
root |
1.72 |
for a seed node as it binds on all local addresses on a fixed port (4040, |
313 |
|
|
customary for aemp). |
314 |
|
|
|
315 |
|
|
# use the aemp commandline utility |
316 |
root |
1.122 |
# aemp profile seed binds '*:4040' |
317 |
root |
1.72 |
|
318 |
|
|
# then use it |
319 |
|
|
configure profile => "seed"; |
320 |
root |
1.34 |
|
321 |
root |
1.72 |
# or simply use aemp from the shell again: |
322 |
|
|
# aemp run profile seed |
323 |
root |
1.34 |
|
324 |
root |
1.72 |
# or provide a nicer-to-remember nodeid |
325 |
|
|
# aemp run profile seed nodeid "$(hostname)" |
326 |
root |
1.34 |
|
327 |
root |
1.22 |
=item $SELF |
328 |
|
|
|
329 |
|
|
Contains the current port id while executing C<rcv> callbacks or C<psub> |
330 |
|
|
blocks. |
331 |
root |
1.3 |
|
332 |
root |
1.67 |
=item *SELF, SELF, %SELF, @SELF... |
333 |
root |
1.22 |
|
334 |
|
|
Due to some quirks in how perl exports variables, it is impossible to |
335 |
root |
1.67 |
just export C<$SELF>, all the symbols named C<SELF> are exported by this |
336 |
root |
1.22 |
module, but only C<$SELF> is currently used. |
337 |
root |
1.3 |
|
338 |
root |
1.33 |
=item snd $port, type => @data |
339 |
root |
1.3 |
|
340 |
root |
1.33 |
=item snd $port, @msg |
341 |
root |
1.3 |
|
342 |
root |
1.67 |
Send the given message to the given port, which can identify either a |
343 |
|
|
local or a remote port, and must be a port ID. |
344 |
root |
1.8 |
|
345 |
root |
1.67 |
While the message can be almost anything, it is highly recommended to |
346 |
|
|
use a string as first element (a port ID, or some word that indicates a |
347 |
|
|
request type etc.) and to consist if only simple perl values (scalars, |
348 |
|
|
arrays, hashes) - if you think you need to pass an object, think again. |
349 |
|
|
|
350 |
|
|
The message data logically becomes read-only after a call to this |
351 |
|
|
function: modifying any argument (or values referenced by them) is |
352 |
|
|
forbidden, as there can be considerable time between the call to C<snd> |
353 |
|
|
and the time the message is actually being serialised - in fact, it might |
354 |
|
|
never be copied as within the same process it is simply handed to the |
355 |
|
|
receiving port. |
356 |
root |
1.3 |
|
357 |
|
|
The type of data you can transfer depends on the transport protocol: when |
358 |
|
|
JSON is used, then only strings, numbers and arrays and hashes consisting |
359 |
|
|
of those are allowed (no objects). When Storable is used, then anything |
360 |
|
|
that Storable can serialise and deserialise is allowed, and for the local |
361 |
root |
1.67 |
node, anything can be passed. Best rely only on the common denominator of |
362 |
|
|
these. |
363 |
root |
1.3 |
|
364 |
root |
1.22 |
=item $local_port = port |
365 |
root |
1.2 |
|
366 |
root |
1.50 |
Create a new local port object and returns its port ID. Initially it has |
367 |
|
|
no callbacks set and will throw an error when it receives messages. |
368 |
root |
1.10 |
|
369 |
root |
1.50 |
=item $local_port = port { my @msg = @_ } |
370 |
root |
1.15 |
|
371 |
root |
1.50 |
Creates a new local port, and returns its ID. Semantically the same as |
372 |
|
|
creating a port and calling C<rcv $port, $callback> on it. |
373 |
root |
1.15 |
|
374 |
root |
1.50 |
The block will be called for every message received on the port, with the |
375 |
|
|
global variable C<$SELF> set to the port ID. Runtime errors will cause the |
376 |
|
|
port to be C<kil>ed. The message will be passed as-is, no extra argument |
377 |
|
|
(i.e. no port ID) will be passed to the callback. |
378 |
root |
1.15 |
|
379 |
root |
1.50 |
If you want to stop/destroy the port, simply C<kil> it: |
380 |
root |
1.15 |
|
381 |
root |
1.50 |
my $port = port { |
382 |
|
|
my @msg = @_; |
383 |
|
|
... |
384 |
|
|
kil $SELF; |
385 |
root |
1.15 |
}; |
386 |
root |
1.10 |
|
387 |
|
|
=cut |
388 |
|
|
|
389 |
root |
1.33 |
sub rcv($@); |
390 |
|
|
|
391 |
root |
1.50 |
sub _kilme { |
392 |
|
|
die "received message on port without callback"; |
393 |
|
|
} |
394 |
|
|
|
395 |
root |
1.22 |
sub port(;&) { |
396 |
root |
1.123 |
my $id = $UNIQ . ++$ID; |
397 |
root |
1.22 |
my $port = "$NODE#$id"; |
398 |
|
|
|
399 |
root |
1.50 |
rcv $port, shift || \&_kilme; |
400 |
root |
1.10 |
|
401 |
root |
1.22 |
$port |
402 |
root |
1.10 |
} |
403 |
|
|
|
404 |
root |
1.50 |
=item rcv $local_port, $callback->(@msg) |
405 |
root |
1.31 |
|
406 |
root |
1.50 |
Replaces the default callback on the specified port. There is no way to |
407 |
|
|
remove the default callback: use C<sub { }> to disable it, or better |
408 |
|
|
C<kil> the port when it is no longer needed. |
409 |
root |
1.3 |
|
410 |
root |
1.33 |
The global C<$SELF> (exported by this module) contains C<$port> while |
411 |
root |
1.50 |
executing the callback. Runtime errors during callback execution will |
412 |
|
|
result in the port being C<kil>ed. |
413 |
root |
1.22 |
|
414 |
root |
1.50 |
The default callback received all messages not matched by a more specific |
415 |
|
|
C<tag> match. |
416 |
root |
1.22 |
|
417 |
root |
1.50 |
=item rcv $local_port, tag => $callback->(@msg_without_tag), ... |
418 |
root |
1.3 |
|
419 |
root |
1.54 |
Register (or replace) callbacks to be called on messages starting with the |
420 |
|
|
given tag on the given port (and return the port), or unregister it (when |
421 |
|
|
C<$callback> is C<$undef> or missing). There can only be one callback |
422 |
|
|
registered for each tag. |
423 |
root |
1.3 |
|
424 |
root |
1.50 |
The original message will be passed to the callback, after the first |
425 |
|
|
element (the tag) has been removed. The callback will use the same |
426 |
|
|
environment as the default callback (see above). |
427 |
root |
1.3 |
|
428 |
root |
1.36 |
Example: create a port and bind receivers on it in one go. |
429 |
|
|
|
430 |
|
|
my $port = rcv port, |
431 |
root |
1.50 |
msg1 => sub { ... }, |
432 |
|
|
msg2 => sub { ... }, |
433 |
root |
1.36 |
; |
434 |
|
|
|
435 |
|
|
Example: create a port, bind receivers and send it in a message elsewhere |
436 |
|
|
in one go: |
437 |
|
|
|
438 |
|
|
snd $otherport, reply => |
439 |
|
|
rcv port, |
440 |
root |
1.50 |
msg1 => sub { ... }, |
441 |
root |
1.36 |
... |
442 |
|
|
; |
443 |
|
|
|
444 |
root |
1.54 |
Example: temporarily register a rcv callback for a tag matching some port |
445 |
root |
1.102 |
(e.g. for an rpc reply) and unregister it after a message was received. |
446 |
root |
1.54 |
|
447 |
|
|
rcv $port, $otherport => sub { |
448 |
|
|
my @reply = @_; |
449 |
|
|
|
450 |
|
|
rcv $SELF, $otherport; |
451 |
|
|
}; |
452 |
|
|
|
453 |
root |
1.3 |
=cut |
454 |
|
|
|
455 |
|
|
sub rcv($@) { |
456 |
root |
1.33 |
my $port = shift; |
457 |
root |
1.75 |
my ($nodeid, $portid) = split /#/, $port, 2; |
458 |
root |
1.3 |
|
459 |
root |
1.75 |
$NODE{$nodeid} == $NODE{""} |
460 |
root |
1.33 |
or Carp::croak "$port: rcv can only be called on local ports, caught"; |
461 |
root |
1.22 |
|
462 |
root |
1.50 |
while (@_) { |
463 |
|
|
if (ref $_[0]) { |
464 |
|
|
if (my $self = $PORT_DATA{$portid}) { |
465 |
|
|
"AnyEvent::MP::Port" eq ref $self |
466 |
|
|
or Carp::croak "$port: rcv can only be called on message matching ports, caught"; |
467 |
root |
1.33 |
|
468 |
root |
1.103 |
$self->[0] = shift; |
469 |
root |
1.50 |
} else { |
470 |
|
|
my $cb = shift; |
471 |
|
|
$PORT{$portid} = sub { |
472 |
|
|
local $SELF = $port; |
473 |
|
|
eval { &$cb }; _self_die if $@; |
474 |
|
|
}; |
475 |
|
|
} |
476 |
|
|
} elsif (defined $_[0]) { |
477 |
|
|
my $self = $PORT_DATA{$portid} ||= do { |
478 |
root |
1.103 |
my $self = bless [$PORT{$portid} || sub { }, { }, $port], "AnyEvent::MP::Port"; |
479 |
root |
1.50 |
|
480 |
|
|
$PORT{$portid} = sub { |
481 |
|
|
local $SELF = $port; |
482 |
|
|
|
483 |
|
|
if (my $cb = $self->[1]{$_[0]}) { |
484 |
|
|
shift; |
485 |
|
|
eval { &$cb }; _self_die if $@; |
486 |
|
|
} else { |
487 |
|
|
&{ $self->[0] }; |
488 |
root |
1.33 |
} |
489 |
|
|
}; |
490 |
root |
1.50 |
|
491 |
|
|
$self |
492 |
root |
1.33 |
}; |
493 |
|
|
|
494 |
root |
1.50 |
"AnyEvent::MP::Port" eq ref $self |
495 |
|
|
or Carp::croak "$port: rcv can only be called on message matching ports, caught"; |
496 |
root |
1.22 |
|
497 |
root |
1.50 |
my ($tag, $cb) = splice @_, 0, 2; |
498 |
root |
1.33 |
|
499 |
root |
1.50 |
if (defined $cb) { |
500 |
|
|
$self->[1]{$tag} = $cb; |
501 |
root |
1.33 |
} else { |
502 |
root |
1.50 |
delete $self->[1]{$tag}; |
503 |
root |
1.33 |
} |
504 |
root |
1.22 |
} |
505 |
root |
1.3 |
} |
506 |
root |
1.31 |
|
507 |
root |
1.33 |
$port |
508 |
root |
1.2 |
} |
509 |
|
|
|
510 |
root |
1.101 |
=item peval $port, $coderef[, @args] |
511 |
|
|
|
512 |
|
|
Evaluates the given C<$codref> within the contetx of C<$port>, that is, |
513 |
|
|
when the code throews an exception the C<$port> will be killed. |
514 |
|
|
|
515 |
|
|
Any remaining args will be passed to the callback. Any return values will |
516 |
|
|
be returned to the caller. |
517 |
|
|
|
518 |
|
|
This is useful when you temporarily want to execute code in the context of |
519 |
|
|
a port. |
520 |
|
|
|
521 |
|
|
Example: create a port and run some initialisation code in it's context. |
522 |
|
|
|
523 |
|
|
my $port = port { ... }; |
524 |
|
|
|
525 |
|
|
peval $port, sub { |
526 |
|
|
init |
527 |
|
|
or die "unable to init"; |
528 |
|
|
}; |
529 |
|
|
|
530 |
|
|
=cut |
531 |
|
|
|
532 |
|
|
sub peval($$) { |
533 |
|
|
local $SELF = shift; |
534 |
|
|
my $cb = shift; |
535 |
|
|
|
536 |
|
|
if (wantarray) { |
537 |
|
|
my @res = eval { &$cb }; |
538 |
|
|
_self_die if $@; |
539 |
|
|
@res |
540 |
|
|
} else { |
541 |
|
|
my $res = eval { &$cb }; |
542 |
|
|
_self_die if $@; |
543 |
|
|
$res |
544 |
|
|
} |
545 |
|
|
} |
546 |
|
|
|
547 |
root |
1.22 |
=item $closure = psub { BLOCK } |
548 |
root |
1.2 |
|
549 |
root |
1.22 |
Remembers C<$SELF> and creates a closure out of the BLOCK. When the |
550 |
|
|
closure is executed, sets up the environment in the same way as in C<rcv> |
551 |
|
|
callbacks, i.e. runtime errors will cause the port to get C<kil>ed. |
552 |
|
|
|
553 |
root |
1.101 |
The effect is basically as if it returned C<< sub { peval $SELF, sub { |
554 |
root |
1.114 |
BLOCK }, @_ } >>. |
555 |
root |
1.101 |
|
556 |
root |
1.22 |
This is useful when you register callbacks from C<rcv> callbacks: |
557 |
|
|
|
558 |
|
|
rcv delayed_reply => sub { |
559 |
|
|
my ($delay, @reply) = @_; |
560 |
|
|
my $timer = AE::timer $delay, 0, psub { |
561 |
|
|
snd @reply, $SELF; |
562 |
|
|
}; |
563 |
|
|
}; |
564 |
root |
1.3 |
|
565 |
root |
1.8 |
=cut |
566 |
root |
1.3 |
|
567 |
root |
1.22 |
sub psub(&) { |
568 |
|
|
my $cb = shift; |
569 |
root |
1.3 |
|
570 |
root |
1.22 |
my $port = $SELF |
571 |
|
|
or Carp::croak "psub can only be called from within rcv or psub callbacks, not"; |
572 |
root |
1.1 |
|
573 |
root |
1.22 |
sub { |
574 |
|
|
local $SELF = $port; |
575 |
root |
1.2 |
|
576 |
root |
1.22 |
if (wantarray) { |
577 |
|
|
my @res = eval { &$cb }; |
578 |
|
|
_self_die if $@; |
579 |
|
|
@res |
580 |
|
|
} else { |
581 |
|
|
my $res = eval { &$cb }; |
582 |
|
|
_self_die if $@; |
583 |
|
|
$res |
584 |
|
|
} |
585 |
|
|
} |
586 |
root |
1.2 |
} |
587 |
|
|
|
588 |
root |
1.67 |
=item $guard = mon $port, $cb->(@reason) # call $cb when $port dies |
589 |
root |
1.32 |
|
590 |
root |
1.67 |
=item $guard = mon $port, $rcvport # kill $rcvport when $port dies |
591 |
root |
1.36 |
|
592 |
root |
1.67 |
=item $guard = mon $port # kill $SELF when $port dies |
593 |
root |
1.32 |
|
594 |
root |
1.67 |
=item $guard = mon $port, $rcvport, @msg # send a message when $port dies |
595 |
root |
1.32 |
|
596 |
root |
1.42 |
Monitor the given port and do something when the port is killed or |
597 |
|
|
messages to it were lost, and optionally return a guard that can be used |
598 |
|
|
to stop monitoring again. |
599 |
|
|
|
600 |
root |
1.36 |
In the first form (callback), the callback is simply called with any |
601 |
|
|
number of C<@reason> elements (no @reason means that the port was deleted |
602 |
root |
1.32 |
"normally"). Note also that I<< the callback B<must> never die >>, so use |
603 |
|
|
C<eval> if unsure. |
604 |
|
|
|
605 |
root |
1.43 |
In the second form (another port given), the other port (C<$rcvport>) |
606 |
elmex |
1.77 |
will be C<kil>'ed with C<@reason>, if a @reason was specified, i.e. on |
607 |
root |
1.36 |
"normal" kils nothing happens, while under all other conditions, the other |
608 |
|
|
port is killed with the same reason. |
609 |
root |
1.32 |
|
610 |
root |
1.36 |
The third form (kill self) is the same as the second form, except that |
611 |
|
|
C<$rvport> defaults to C<$SELF>. |
612 |
|
|
|
613 |
|
|
In the last form (message), a message of the form C<@msg, @reason> will be |
614 |
|
|
C<snd>. |
615 |
root |
1.32 |
|
616 |
root |
1.79 |
Monitoring-actions are one-shot: once messages are lost (and a monitoring |
617 |
|
|
alert was raised), they are removed and will not trigger again. |
618 |
|
|
|
619 |
root |
1.37 |
As a rule of thumb, monitoring requests should always monitor a port from |
620 |
|
|
a local port (or callback). The reason is that kill messages might get |
621 |
|
|
lost, just like any other message. Another less obvious reason is that |
622 |
elmex |
1.77 |
even monitoring requests can get lost (for example, when the connection |
623 |
root |
1.37 |
to the other node goes down permanently). When monitoring a port locally |
624 |
|
|
these problems do not exist. |
625 |
|
|
|
626 |
root |
1.79 |
C<mon> effectively guarantees that, in the absence of hardware failures, |
627 |
|
|
after starting the monitor, either all messages sent to the port will |
628 |
|
|
arrive, or the monitoring action will be invoked after possible message |
629 |
|
|
loss has been detected. No messages will be lost "in between" (after |
630 |
|
|
the first lost message no further messages will be received by the |
631 |
|
|
port). After the monitoring action was invoked, further messages might get |
632 |
|
|
delivered again. |
633 |
|
|
|
634 |
|
|
Inter-host-connection timeouts and monitoring depend on the transport |
635 |
|
|
used. The only transport currently implemented is TCP, and AnyEvent::MP |
636 |
|
|
relies on TCP to detect node-downs (this can take 10-15 minutes on a |
637 |
elmex |
1.96 |
non-idle connection, and usually around two hours for idle connections). |
638 |
root |
1.79 |
|
639 |
|
|
This means that monitoring is good for program errors and cleaning up |
640 |
|
|
stuff eventually, but they are no replacement for a timeout when you need |
641 |
|
|
to ensure some maximum latency. |
642 |
|
|
|
643 |
root |
1.32 |
Example: call a given callback when C<$port> is killed. |
644 |
|
|
|
645 |
|
|
mon $port, sub { warn "port died because of <@_>\n" }; |
646 |
|
|
|
647 |
|
|
Example: kill ourselves when C<$port> is killed abnormally. |
648 |
|
|
|
649 |
root |
1.36 |
mon $port; |
650 |
root |
1.32 |
|
651 |
root |
1.36 |
Example: send us a restart message when another C<$port> is killed. |
652 |
root |
1.32 |
|
653 |
|
|
mon $port, $self => "restart"; |
654 |
|
|
|
655 |
|
|
=cut |
656 |
|
|
|
657 |
|
|
sub mon { |
658 |
root |
1.75 |
my ($nodeid, $port) = split /#/, shift, 2; |
659 |
root |
1.32 |
|
660 |
root |
1.75 |
my $node = $NODE{$nodeid} || add_node $nodeid; |
661 |
root |
1.32 |
|
662 |
root |
1.41 |
my $cb = @_ ? shift : $SELF || Carp::croak 'mon: called with one argument only, but $SELF not set,'; |
663 |
root |
1.32 |
|
664 |
|
|
unless (ref $cb) { |
665 |
|
|
if (@_) { |
666 |
|
|
# send a kill info message |
667 |
root |
1.41 |
my (@msg) = ($cb, @_); |
668 |
root |
1.32 |
$cb = sub { snd @msg, @_ }; |
669 |
|
|
} else { |
670 |
|
|
# simply kill other port |
671 |
|
|
my $port = $cb; |
672 |
|
|
$cb = sub { kil $port, @_ if @_ }; |
673 |
|
|
} |
674 |
|
|
} |
675 |
|
|
|
676 |
|
|
$node->monitor ($port, $cb); |
677 |
|
|
|
678 |
|
|
defined wantarray |
679 |
root |
1.124 |
and ($cb += 0, Guard::guard { $node->unmonitor ($port, $cb) }) |
680 |
root |
1.32 |
} |
681 |
|
|
|
682 |
|
|
=item $guard = mon_guard $port, $ref, $ref... |
683 |
|
|
|
684 |
|
|
Monitors the given C<$port> and keeps the passed references. When the port |
685 |
|
|
is killed, the references will be freed. |
686 |
|
|
|
687 |
|
|
Optionally returns a guard that will stop the monitoring. |
688 |
|
|
|
689 |
|
|
This function is useful when you create e.g. timers or other watchers and |
690 |
root |
1.67 |
want to free them when the port gets killed (note the use of C<psub>): |
691 |
root |
1.32 |
|
692 |
|
|
$port->rcv (start => sub { |
693 |
root |
1.67 |
my $timer; $timer = mon_guard $port, AE::timer 1, 1, psub { |
694 |
root |
1.32 |
undef $timer if 0.9 < rand; |
695 |
|
|
}); |
696 |
|
|
}); |
697 |
|
|
|
698 |
|
|
=cut |
699 |
|
|
|
700 |
|
|
sub mon_guard { |
701 |
|
|
my ($port, @refs) = @_; |
702 |
|
|
|
703 |
root |
1.36 |
#TODO: mon-less form? |
704 |
|
|
|
705 |
root |
1.32 |
mon $port, sub { 0 && @refs } |
706 |
|
|
} |
707 |
|
|
|
708 |
root |
1.33 |
=item kil $port[, @reason] |
709 |
root |
1.32 |
|
710 |
|
|
Kill the specified port with the given C<@reason>. |
711 |
|
|
|
712 |
root |
1.107 |
If no C<@reason> is specified, then the port is killed "normally" - |
713 |
|
|
monitor callback will be invoked, but the kil will not cause linked ports |
714 |
|
|
(C<mon $mport, $lport> form) to get killed. |
715 |
root |
1.32 |
|
716 |
root |
1.107 |
If a C<@reason> is specified, then linked ports (C<mon $mport, $lport> |
717 |
|
|
form) get killed with the same reason. |
718 |
root |
1.32 |
|
719 |
|
|
Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks |
720 |
|
|
will be reported as reason C<< die => $@ >>. |
721 |
|
|
|
722 |
|
|
Transport/communication errors are reported as C<< transport_error => |
723 |
|
|
$message >>. |
724 |
|
|
|
725 |
root |
1.38 |
=cut |
726 |
|
|
|
727 |
|
|
=item $port = spawn $node, $initfunc[, @initdata] |
728 |
|
|
|
729 |
|
|
Creates a port on the node C<$node> (which can also be a port ID, in which |
730 |
|
|
case it's the node where that port resides). |
731 |
|
|
|
732 |
root |
1.67 |
The port ID of the newly created port is returned immediately, and it is |
733 |
|
|
possible to immediately start sending messages or to monitor the port. |
734 |
root |
1.38 |
|
735 |
root |
1.67 |
After the port has been created, the init function is called on the remote |
736 |
|
|
node, in the same context as a C<rcv> callback. This function must be a |
737 |
|
|
fully-qualified function name (e.g. C<MyApp::Chat::Server::init>). To |
738 |
|
|
specify a function in the main program, use C<::name>. |
739 |
root |
1.38 |
|
740 |
|
|
If the function doesn't exist, then the node tries to C<require> |
741 |
|
|
the package, then the package above the package and so on (e.g. |
742 |
|
|
C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function |
743 |
|
|
exists or it runs out of package names. |
744 |
|
|
|
745 |
|
|
The init function is then called with the newly-created port as context |
746 |
root |
1.82 |
object (C<$SELF>) and the C<@initdata> values as arguments. It I<must> |
747 |
|
|
call one of the C<rcv> functions to set callbacks on C<$SELF>, otherwise |
748 |
|
|
the port might not get created. |
749 |
root |
1.38 |
|
750 |
root |
1.67 |
A common idiom is to pass a local port, immediately monitor the spawned |
751 |
|
|
port, and in the remote init function, immediately monitor the passed |
752 |
|
|
local port. This two-way monitoring ensures that both ports get cleaned up |
753 |
|
|
when there is a problem. |
754 |
root |
1.38 |
|
755 |
root |
1.80 |
C<spawn> guarantees that the C<$initfunc> has no visible effects on the |
756 |
|
|
caller before C<spawn> returns (by delaying invocation when spawn is |
757 |
|
|
called for the local node). |
758 |
|
|
|
759 |
root |
1.38 |
Example: spawn a chat server port on C<$othernode>. |
760 |
|
|
|
761 |
|
|
# this node, executed from within a port context: |
762 |
|
|
my $server = spawn $othernode, "MyApp::Chat::Server::connect", $SELF; |
763 |
|
|
mon $server; |
764 |
|
|
|
765 |
|
|
# init function on C<$othernode> |
766 |
|
|
sub connect { |
767 |
|
|
my ($srcport) = @_; |
768 |
|
|
|
769 |
|
|
mon $srcport; |
770 |
|
|
|
771 |
|
|
rcv $SELF, sub { |
772 |
|
|
... |
773 |
|
|
}; |
774 |
|
|
} |
775 |
|
|
|
776 |
|
|
=cut |
777 |
|
|
|
778 |
|
|
sub _spawn { |
779 |
|
|
my $port = shift; |
780 |
|
|
my $init = shift; |
781 |
|
|
|
782 |
root |
1.82 |
# rcv will create the actual port |
783 |
root |
1.38 |
local $SELF = "$NODE#$port"; |
784 |
|
|
eval { |
785 |
|
|
&{ load_func $init } |
786 |
|
|
}; |
787 |
|
|
_self_die if $@; |
788 |
|
|
} |
789 |
|
|
|
790 |
|
|
sub spawn(@) { |
791 |
root |
1.75 |
my ($nodeid, undef) = split /#/, shift, 2; |
792 |
root |
1.38 |
|
793 |
root |
1.123 |
my $id = $RUNIQ . ++$ID; |
794 |
root |
1.38 |
|
795 |
root |
1.39 |
$_[0] =~ /::/ |
796 |
|
|
or Carp::croak "spawn init function must be a fully-qualified name, caught"; |
797 |
|
|
|
798 |
root |
1.75 |
snd_to_func $nodeid, "AnyEvent::MP::_spawn" => $id, @_; |
799 |
root |
1.38 |
|
800 |
root |
1.75 |
"$nodeid#$id" |
801 |
root |
1.38 |
} |
802 |
|
|
|
803 |
root |
1.121 |
|
804 |
root |
1.59 |
=item after $timeout, @msg |
805 |
|
|
|
806 |
|
|
=item after $timeout, $callback |
807 |
|
|
|
808 |
|
|
Either sends the given message, or call the given callback, after the |
809 |
|
|
specified number of seconds. |
810 |
|
|
|
811 |
root |
1.67 |
This is simply a utility function that comes in handy at times - the |
812 |
|
|
AnyEvent::MP author is not convinced of the wisdom of having it, though, |
813 |
|
|
so it may go away in the future. |
814 |
root |
1.59 |
|
815 |
|
|
=cut |
816 |
|
|
|
817 |
|
|
sub after($@) { |
818 |
|
|
my ($timeout, @action) = @_; |
819 |
|
|
|
820 |
|
|
my $t; $t = AE::timer $timeout, 0, sub { |
821 |
|
|
undef $t; |
822 |
|
|
ref $action[0] |
823 |
|
|
? $action[0]() |
824 |
|
|
: snd @action; |
825 |
|
|
}; |
826 |
|
|
} |
827 |
|
|
|
828 |
root |
1.87 |
=item cal $port, @msg, $callback[, $timeout] |
829 |
|
|
|
830 |
|
|
A simple form of RPC - sends a message to the given C<$port> with the |
831 |
|
|
given contents (C<@msg>), but adds a reply port to the message. |
832 |
|
|
|
833 |
|
|
The reply port is created temporarily just for the purpose of receiving |
834 |
|
|
the reply, and will be C<kil>ed when no longer needed. |
835 |
|
|
|
836 |
|
|
A reply message sent to the port is passed to the C<$callback> as-is. |
837 |
|
|
|
838 |
|
|
If an optional time-out (in seconds) is given and it is not C<undef>, |
839 |
|
|
then the callback will be called without any arguments after the time-out |
840 |
|
|
elapsed and the port is C<kil>ed. |
841 |
|
|
|
842 |
root |
1.98 |
If no time-out is given (or it is C<undef>), then the local port will |
843 |
|
|
monitor the remote port instead, so it eventually gets cleaned-up. |
844 |
root |
1.87 |
|
845 |
|
|
Currently this function returns the temporary port, but this "feature" |
846 |
|
|
might go in future versions unless you can make a convincing case that |
847 |
|
|
this is indeed useful for something. |
848 |
|
|
|
849 |
|
|
=cut |
850 |
|
|
|
851 |
|
|
sub cal(@) { |
852 |
|
|
my $timeout = ref $_[-1] ? undef : pop; |
853 |
|
|
my $cb = pop; |
854 |
|
|
|
855 |
|
|
my $port = port { |
856 |
|
|
undef $timeout; |
857 |
|
|
kil $SELF; |
858 |
|
|
&$cb; |
859 |
|
|
}; |
860 |
|
|
|
861 |
|
|
if (defined $timeout) { |
862 |
|
|
$timeout = AE::timer $timeout, 0, sub { |
863 |
|
|
undef $timeout; |
864 |
|
|
kil $port; |
865 |
|
|
$cb->(); |
866 |
|
|
}; |
867 |
|
|
} else { |
868 |
|
|
mon $_[0], sub { |
869 |
|
|
kil $port; |
870 |
|
|
$cb->(); |
871 |
|
|
}; |
872 |
|
|
} |
873 |
|
|
|
874 |
|
|
push @_, $port; |
875 |
|
|
&snd; |
876 |
|
|
|
877 |
|
|
$port |
878 |
|
|
} |
879 |
|
|
|
880 |
root |
1.8 |
=back |
881 |
|
|
|
882 |
root |
1.124 |
=head1 DISTRIBUTED DATABASE |
883 |
|
|
|
884 |
|
|
AnyEvent::MP comes with a simple distributed database. The database will |
885 |
|
|
be mirrored asynchronously at all global nodes. Other nodes bind to one of |
886 |
|
|
the global nodes for their needs. |
887 |
|
|
|
888 |
|
|
The database consists of a two-level hash - a hash contains a hash which |
889 |
|
|
contains values. |
890 |
|
|
|
891 |
|
|
The top level hash key is called "family", and the second-level hash key |
892 |
|
|
is simply called "key". |
893 |
|
|
|
894 |
root |
1.125 |
The family must be alphanumeric, i.e. start with a letter and consist |
895 |
|
|
of letters, digits, underscores and colons (C<[A-Za-z][A-Za-z0-9_:]*>, |
896 |
|
|
pretty much like Perl module names. |
897 |
root |
1.124 |
|
898 |
root |
1.125 |
As the family namespace is global, it is recommended to prefix family names |
899 |
root |
1.124 |
with the name of the application or module using it. |
900 |
|
|
|
901 |
root |
1.125 |
The keys must be strings, with no other limitations. |
902 |
|
|
|
903 |
root |
1.124 |
The values should preferably be strings, but other perl scalars should |
904 |
root |
1.125 |
work as well (such as undef, arrays and hashes). |
905 |
root |
1.124 |
|
906 |
|
|
Every database entry is owned by one node - adding the same family/key |
907 |
|
|
combination on multiple nodes will not cause discomfort for AnyEvent::MP, |
908 |
|
|
but the result might be nondeterministic, i.e. the key might have |
909 |
|
|
different values on different nodes. |
910 |
|
|
|
911 |
|
|
=item db_set $family => $key => $value |
912 |
|
|
|
913 |
|
|
Sets (or replaces) a key to the database. |
914 |
|
|
|
915 |
|
|
=item db_del $family => $key |
916 |
|
|
|
917 |
|
|
Deletes a key from the database. |
918 |
|
|
|
919 |
|
|
=item $guard = db_reg $family => $key [=> $value] |
920 |
|
|
|
921 |
|
|
Sets the key on the database and returns a guard. When the guard is |
922 |
|
|
destroyed, the key is deleted from the database. If C<$value> is missing, |
923 |
|
|
then C<undef> is used. |
924 |
|
|
|
925 |
|
|
=cut |
926 |
|
|
|
927 |
|
|
=back |
928 |
|
|
|
929 |
root |
1.26 |
=head1 AnyEvent::MP vs. Distributed Erlang |
930 |
|
|
|
931 |
root |
1.35 |
AnyEvent::MP got lots of its ideas from distributed Erlang (Erlang node |
932 |
|
|
== aemp node, Erlang process == aemp port), so many of the documents and |
933 |
|
|
programming techniques employed by Erlang apply to AnyEvent::MP. Here is a |
934 |
root |
1.27 |
sample: |
935 |
|
|
|
936 |
root |
1.95 |
http://www.erlang.se/doc/programming_rules.shtml |
937 |
|
|
http://erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4 |
938 |
|
|
http://erlang.org/download/erlang-book-part1.pdf # chapters 5 and 6 |
939 |
|
|
http://erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5 |
940 |
root |
1.27 |
|
941 |
|
|
Despite the similarities, there are also some important differences: |
942 |
root |
1.26 |
|
943 |
|
|
=over 4 |
944 |
|
|
|
945 |
root |
1.65 |
=item * Node IDs are arbitrary strings in AEMP. |
946 |
root |
1.26 |
|
947 |
root |
1.65 |
Erlang relies on special naming and DNS to work everywhere in the same |
948 |
|
|
way. AEMP relies on each node somehow knowing its own address(es) (e.g. by |
949 |
root |
1.99 |
configuration or DNS), and possibly the addresses of some seed nodes, but |
950 |
|
|
will otherwise discover other nodes (and their IDs) itself. |
951 |
root |
1.27 |
|
952 |
root |
1.54 |
=item * Erlang has a "remote ports are like local ports" philosophy, AEMP |
953 |
root |
1.51 |
uses "local ports are like remote ports". |
954 |
|
|
|
955 |
|
|
The failure modes for local ports are quite different (runtime errors |
956 |
|
|
only) then for remote ports - when a local port dies, you I<know> it dies, |
957 |
|
|
when a connection to another node dies, you know nothing about the other |
958 |
|
|
port. |
959 |
|
|
|
960 |
|
|
Erlang pretends remote ports are as reliable as local ports, even when |
961 |
|
|
they are not. |
962 |
|
|
|
963 |
|
|
AEMP encourages a "treat remote ports differently" philosophy, with local |
964 |
|
|
ports being the special case/exception, where transport errors cannot |
965 |
|
|
occur. |
966 |
|
|
|
967 |
root |
1.26 |
=item * Erlang uses processes and a mailbox, AEMP does not queue. |
968 |
|
|
|
969 |
root |
1.119 |
Erlang uses processes that selectively receive messages out of order, and |
970 |
|
|
therefore needs a queue. AEMP is event based, queuing messages would serve |
971 |
|
|
no useful purpose. For the same reason the pattern-matching abilities |
972 |
|
|
of AnyEvent::MP are more limited, as there is little need to be able to |
973 |
elmex |
1.77 |
filter messages without dequeuing them. |
974 |
root |
1.26 |
|
975 |
root |
1.119 |
This is not a philosophical difference, but simply stems from AnyEvent::MP |
976 |
|
|
being event-based, while Erlang is process-based. |
977 |
|
|
|
978 |
|
|
You cna have a look at L<Coro::MP> for a more Erlang-like process model on |
979 |
|
|
top of AEMP and Coro threads. |
980 |
root |
1.26 |
|
981 |
|
|
=item * Erlang sends are synchronous, AEMP sends are asynchronous. |
982 |
|
|
|
983 |
root |
1.119 |
Sending messages in Erlang is synchronous and blocks the process until |
984 |
|
|
a conenction has been established and the message sent (and so does not |
985 |
|
|
need a queue that can overflow). AEMP sends return immediately, connection |
986 |
|
|
establishment is handled in the background. |
987 |
root |
1.26 |
|
988 |
root |
1.51 |
=item * Erlang suffers from silent message loss, AEMP does not. |
989 |
root |
1.26 |
|
990 |
root |
1.99 |
Erlang implements few guarantees on messages delivery - messages can get |
991 |
|
|
lost without any of the processes realising it (i.e. you send messages a, |
992 |
|
|
b, and c, and the other side only receives messages a and c). |
993 |
root |
1.26 |
|
994 |
root |
1.117 |
AEMP guarantees (modulo hardware errors) correct ordering, and the |
995 |
|
|
guarantee that after one message is lost, all following ones sent to the |
996 |
|
|
same port are lost as well, until monitoring raises an error, so there are |
997 |
|
|
no silent "holes" in the message sequence. |
998 |
root |
1.26 |
|
999 |
root |
1.119 |
If you want your software to be very reliable, you have to cope with |
1000 |
|
|
corrupted and even out-of-order messages in both Erlang and AEMP. AEMP |
1001 |
|
|
simply tries to work better in common error cases, such as when a network |
1002 |
|
|
link goes down. |
1003 |
|
|
|
1004 |
root |
1.26 |
=item * Erlang can send messages to the wrong port, AEMP does not. |
1005 |
|
|
|
1006 |
root |
1.119 |
In Erlang it is quite likely that a node that restarts reuses an Erlang |
1007 |
|
|
process ID known to other nodes for a completely different process, |
1008 |
|
|
causing messages destined for that process to end up in an unrelated |
1009 |
|
|
process. |
1010 |
root |
1.26 |
|
1011 |
root |
1.119 |
AEMP does not reuse port IDs, so old messages or old port IDs floating |
1012 |
root |
1.26 |
around in the network will not be sent to an unrelated port. |
1013 |
|
|
|
1014 |
|
|
=item * Erlang uses unprotected connections, AEMP uses secure |
1015 |
|
|
authentication and can use TLS. |
1016 |
|
|
|
1017 |
root |
1.66 |
AEMP can use a proven protocol - TLS - to protect connections and |
1018 |
root |
1.26 |
securely authenticate nodes. |
1019 |
|
|
|
1020 |
root |
1.28 |
=item * The AEMP protocol is optimised for both text-based and binary |
1021 |
|
|
communications. |
1022 |
|
|
|
1023 |
root |
1.66 |
The AEMP protocol, unlike the Erlang protocol, supports both programming |
1024 |
root |
1.119 |
language independent text-only protocols (good for debugging), and binary, |
1025 |
root |
1.67 |
language-specific serialisers (e.g. Storable). By default, unless TLS is |
1026 |
|
|
used, the protocol is actually completely text-based. |
1027 |
root |
1.28 |
|
1028 |
|
|
It has also been carefully designed to be implementable in other languages |
1029 |
root |
1.66 |
with a minimum of work while gracefully degrading functionality to make the |
1030 |
root |
1.28 |
protocol simple. |
1031 |
|
|
|
1032 |
root |
1.35 |
=item * AEMP has more flexible monitoring options than Erlang. |
1033 |
|
|
|
1034 |
root |
1.119 |
In Erlang, you can chose to receive I<all> exit signals as messages or |
1035 |
|
|
I<none>, there is no in-between, so monitoring single Erlang processes is |
1036 |
|
|
difficult to implement. |
1037 |
|
|
|
1038 |
|
|
Monitoring in AEMP is more flexible than in Erlang, as one can choose |
1039 |
|
|
between automatic kill, exit message or callback on a per-port basis. |
1040 |
root |
1.35 |
|
1041 |
root |
1.37 |
=item * Erlang tries to hide remote/local connections, AEMP does not. |
1042 |
root |
1.35 |
|
1043 |
root |
1.67 |
Monitoring in Erlang is not an indicator of process death/crashes, in the |
1044 |
|
|
same way as linking is (except linking is unreliable in Erlang). |
1045 |
root |
1.37 |
|
1046 |
|
|
In AEMP, you don't "look up" registered port names or send to named ports |
1047 |
|
|
that might or might not be persistent. Instead, you normally spawn a port |
1048 |
root |
1.67 |
on the remote node. The init function monitors you, and you monitor the |
1049 |
|
|
remote port. Since both monitors are local to the node, they are much more |
1050 |
|
|
reliable (no need for C<spawn_link>). |
1051 |
root |
1.37 |
|
1052 |
|
|
This also saves round-trips and avoids sending messages to the wrong port |
1053 |
|
|
(hard to do in Erlang). |
1054 |
root |
1.35 |
|
1055 |
root |
1.26 |
=back |
1056 |
|
|
|
1057 |
root |
1.46 |
=head1 RATIONALE |
1058 |
|
|
|
1059 |
|
|
=over 4 |
1060 |
|
|
|
1061 |
root |
1.67 |
=item Why strings for port and node IDs, why not objects? |
1062 |
root |
1.46 |
|
1063 |
|
|
We considered "objects", but found that the actual number of methods |
1064 |
root |
1.67 |
that can be called are quite low. Since port and node IDs travel over |
1065 |
root |
1.46 |
the network frequently, the serialising/deserialising would add lots of |
1066 |
root |
1.67 |
overhead, as well as having to keep a proxy object everywhere. |
1067 |
root |
1.46 |
|
1068 |
|
|
Strings can easily be printed, easily serialised etc. and need no special |
1069 |
|
|
procedures to be "valid". |
1070 |
|
|
|
1071 |
root |
1.110 |
And as a result, a port with just a default receiver consists of a single |
1072 |
root |
1.117 |
code reference stored in a global hash - it can't become much cheaper. |
1073 |
root |
1.47 |
|
1074 |
root |
1.67 |
=item Why favour JSON, why not a real serialising format such as Storable? |
1075 |
root |
1.46 |
|
1076 |
|
|
In fact, any AnyEvent::MP node will happily accept Storable as framing |
1077 |
|
|
format, but currently there is no way to make a node use Storable by |
1078 |
root |
1.67 |
default (although all nodes will accept it). |
1079 |
root |
1.46 |
|
1080 |
|
|
The default framing protocol is JSON because a) JSON::XS is many times |
1081 |
|
|
faster for small messages and b) most importantly, after years of |
1082 |
|
|
experience we found that object serialisation is causing more problems |
1083 |
root |
1.67 |
than it solves: Just like function calls, objects simply do not travel |
1084 |
root |
1.46 |
easily over the network, mostly because they will always be a copy, so you |
1085 |
|
|
always have to re-think your design. |
1086 |
|
|
|
1087 |
|
|
Keeping your messages simple, concentrating on data structures rather than |
1088 |
|
|
objects, will keep your messages clean, tidy and efficient. |
1089 |
|
|
|
1090 |
|
|
=back |
1091 |
|
|
|
1092 |
root |
1.1 |
=head1 SEE ALSO |
1093 |
|
|
|
1094 |
root |
1.68 |
L<AnyEvent::MP::Intro> - a gentle introduction. |
1095 |
|
|
|
1096 |
|
|
L<AnyEvent::MP::Kernel> - more, lower-level, stuff. |
1097 |
|
|
|
1098 |
root |
1.113 |
L<AnyEvent::MP::Global> - network maintenance and port groups, to find |
1099 |
root |
1.68 |
your applications. |
1100 |
|
|
|
1101 |
root |
1.105 |
L<AnyEvent::MP::DataConn> - establish data connections between nodes. |
1102 |
|
|
|
1103 |
root |
1.81 |
L<AnyEvent::MP::LogCatcher> - simple service to display log messages from |
1104 |
|
|
all nodes. |
1105 |
|
|
|
1106 |
root |
1.1 |
L<AnyEvent>. |
1107 |
|
|
|
1108 |
|
|
=head1 AUTHOR |
1109 |
|
|
|
1110 |
|
|
Marc Lehmann <schmorp@schmorp.de> |
1111 |
|
|
http://home.schmorp.de/ |
1112 |
|
|
|
1113 |
|
|
=cut |
1114 |
|
|
|
1115 |
|
|
1 |
1116 |
|
|
|